運用免疫演算法求解複雜系統之可靠度最佳化問題

Immune Algorithm Approach for Reliability Optimization with Complex Systems

蘇建銘(Su, Chien-Ming)

指導教授 ： 游文龍

國立虎尾科技大學/管理學院/工業工程與管理研究所/碩士(2008年)

摘要

系統的可靠度以及建構成本在工業系統設計的領域中，是一個重要的課題。隨著系統結構日趨複雜，當系統中有零件發生失效時，導致系統失效的弊害是相當嚴重的。在降低系統失效的機率而提升系統可靠度的同時，成本往往也相對的呈指數增加。因此，決定複雜系統之可靠度參數是一個複雜的最佳化問題。

本研究針對複雜系統之可靠度與成本之議題，當系統有成本限制或者有系統可靠度下限的情況下，提出以免疫演算法來求解最佳可靠度配置與成本最小化之問題。可靠度的最佳化問題已經被研究數十年之久，通常是應用數學模式或啟發式之求解方法來解決此類型問題，而其中的關鍵在於必須滿足問題所有的限制條件，並求得最佳化的解。

本研究提出一個以生物免疫系統的特性為基礎所衍生的方法，稱為免疫演算法(immune algorithm)，這個方法是利用生物免疫系統中，抗體與抗原結合的關係，以及免疫反應中專一性與記憶性的特徵，藉以增加複雜系統可靠度最佳化問題的求解速度。使用免疫演算法得到的結果，與目前文獻上解可靠度最佳化的問題的結果比較，證明本研究所提出的免疫演算法在求解複雜系統的可靠度最佳化問題上，效能較佳。

關鍵字

可靠度 ； 複雜系統 ； 免疫演算法

並列摘要

In the area of industrial system design, high system reliability and low built cost of system are important issues. With the growing complexity of system architecture, the loss caused by system failure is quite more serious. The relative cost often exponentially increased while reducing the probability of system failures and improving system reliability. Therefore, the decision and optimization of complex system reliability parameters is a complex issue.
This thesis is aimed at the reliability and cost optimization issues of a complex system. Immune algorithm is presented to optimize the system reliability and minimize the built cost of a complex system. The problem of system reliability optimization has been studied for decades-long. Mathematical models or heuristic methods are usually utilized to solve this kind of problems.
The immune algorithm is based on the characteristics of biological immune system which employs the relationship between the antibodies and antigens and the principles of immune reaction of the immune system and the characteristics of specificity and memoriability of the immune system to speed up and assure that the feasible solution converges as quickly as possible. To compare the proposed methodology with the presented literature on the problems of reliability optimization, the results of using immune algorithm which is presented by this thesis are better than those by the others.

並列關鍵字

Reliability ； Complex System ； Immune Algorithm

參考文獻

1. 朱迎善、張鐵柱 (2000) 「遺傳算法在非串並聯系統可靠性最優分配中的應用」 ，中華管理學報，一卷，一期，頁13-20

2. Bentley P. J., Corne D. W. (Eds). (2001). Creative Evolutionary Systems, Morgan Kaufmann Pub (to appear).

3. C. Y. Liu, R. L. Bulfin. (1985). “Optimal Allocation of Redundant Components for Large Systems”, IEEE Transactions on Reliability, Vol.R-34, No.3, pp.241-247.

4. Chern M. S. Jan, R. H., Chern R. J. (1991). “Parametric nonlinear integer programming: the right-hand side case”, European J. Operations Research, Vol. 54, No.2, pp.237-255.

5. Chern M. S., Jan, R. H. (1985). “Parametric programming applied to reliability optimization problems”, IEEE Transaction Reliability, Vol. R-34, pp.165-170.